What is in WJEC A-Level Computer Science Unit 4?

Unit 4 Computer Architecture, Data, Communication and Applications is an A2 unit examined by a written paper. It covers hardware and communication (processor architecture, buses, parallel processing, communication hardware), advanced data representation (floating-point numbers, normalisation, fixed point), organisation and structure of data (relational databases, normalisation to third normal form, SQL, big data and data warehousing), low-level programs and program construction (machine code, assembly language, addressing modes), applications of computer science (artificial intelligence, machine learning, automation), and the ethical, legal, social and environmental impact of computing.

How is WJEC A-Level Computer Science Unit 4 examined?

Unit 4 is an A2 written examination of about 2 hours contributing 20 per cent of the full A level. Questions are drawn directly from the specification statements and include calculations on address-bus width and addressable memory, floating-point and normalisation working, normalising a database to third normal form, tracing addressing modes in assembly, defining and contrasting AI and machine learning, and discussing the impacts of computing. Always confirm the exact length and mark total from the current WJEC specification and past papers.

What is the difference between the mantissa and the exponent in a floating-point number?

A floating-point number is stored as a mantissa times two to the power of an exponent. The mantissa (significand) holds the significant digits and fixes the precision: more mantissa bits give more significant figures and smaller rounding error. The exponent holds the power of two that scales the mantissa and fixes the range: more exponent bits allow larger and smaller magnitudes. With a fixed word length there is a trade-off, because giving more bits to one means fewer for the other.

What are the three normal forms in database normalisation?

First normal form (1NF) requires atomic values, no repeating groups and a primary key. Second normal form (2NF) requires the table to be in 1NF and every non-key field to depend on the whole primary key, removing partial dependencies on part of a composite key. Third normal form (3NF) requires the table to be in 2NF with no non-key field depending on another non-key field, removing transitive dependencies. In 3NF every non-key field depends only on the primary key, which removes the redundancy and anomalies that drive normalisation.

How does machine learning differ from traditional programming?

In traditional programming a programmer writes explicit rules that specify exactly how to turn input into output. In machine learning the rules are not written by hand; instead the system learns patterns from training data, building a model that it uses to make predictions or decisions on new, unseen data. This makes machine learning powerful for tasks too complex to specify with explicit rules, such as image or speech recognition, but it depends heavily on the quality of its training data and can be opaque, so human oversight is needed.

WalesComputer Science

WJEC A-Level Computer Science Unit 4 Computer Architecture, Data, Communication and Applications: a complete overview of architecture, floating point, databases, assembly and impact

A deep-dive WJEC A-Level Computer Science guide to Unit 4 Computer Architecture, Data, Communication and Applications. Covers hardware and communication, floating point and normalisation, databases to third normal form and big data, low-level assembly and addressing modes, AI and machine learning, and the ethical, legal, social and environmental impact of computing.

Generated by Claude Opus 4.823 min readWJECUpdated 2026-06-16

Reviewed by: AI editorial process; not yet individually human-reviewed

Jump to a section

What Unit 4 actually demands
Hardware and communication
Data representation and data types
Organisation and structure of data
Programs and program construction
Algorithms and programs: applications
The impact of computer science
How Unit 4 is examined
The six topics, dot point by dot point
For the official specification

What Unit 4 actually demands

Unit 4 Computer Architecture, Data, Communication and Applications is the second A2 written unit. It deepens the hardware, data and communication content and adds applications and impact. The content spans processor architecture and communication hardware, advanced data representation and floating point, databases normalised to third normal form plus big data, low-level assembly programming, AI and machine learning, and the ethical, legal, social and environmental impacts of computing. It rewards confident calculation (address buses, floating point, normalisation), precise low-level reasoning, and balanced evaluation of impact.

This guide walks through the six topics of the unit, then sets out the exam patterns WJEC repeats. Each topic has a matching dot-point page with worked exam questions; this overview ties them together.

Hardware and communication

The processor connects to memory and peripherals over buses: the address bus carries the location (an n-line bus addresses 2 to the power n locations), the data bus carries the contents, the control bus carries signals. Parallel processing speeds up tasks that divide into independent parts, and communication hardware moves data between machines.

Data representation and data types

Real numbers are stored in floating-point form as a mantissa (precision) and exponent (range), usually in two's complement. Normalisation adjusts the mantissa to a standard range with a compensating exponent change, giving a unique representation with maximum precision. Fixed point and the finite-bit limit explain rounding error.

Organisation and structure of data

Relational databases linked by keys are normalised to remove redundancy: 1NF (atomic values, a key), 2NF (no partial dependency), 3NF (no transitive dependency). SQL queries and edits the data. Big data (volume, velocity, variety) and data warehousing extend the model to the modern data landscape.

Programs and program construction

Machine code is binary; assembly language uses mnemonics translated by an assembler. An instruction has an opcode (operation) and an operand (data or reference). Addressing modes interpret the operand: immediate (the value), direct (its address), indirect and indexed. Assembly gives control but is machine-specific.

Algorithms and programs: applications

Artificial intelligence simulates intelligent behaviour. Machine learning infers patterns from training data rather than following hand-written rules; neural networks and expert systems are examples. Automation brings speed, consistency and safety but risks job displacement and over-reliance. All have limits and need oversight.

The impact of computer science

Computing raises ethical (privacy, consent, bias), legal (data protection, computer misuse, copyright), social (employment, the digital divide) and environmental (energy use, e-waste) issues. Responsible computing weighs costs and benefits, follows the law, and seeks to minimise harm.

How Unit 4 is examined

Unit 4 is an A2 written paper of about 2 hours contributing 20 per cent of the full A level. Questions come straight from the specification statements and mix calculation (address buses, floating point, normalisation, addressing modes) with extended discussion (AI, impact). Drill the calculations until automatic, and practise balanced, two-sided impact answers because the discursive questions reward judgement.

The six topics, dot point by dot point

Each topic has a dot-point answer page with worked exam questions and cross-links. Browse them from this unit overview and the subject hub.

For the official specification

WJEC publishes the full specification, past papers and mark schemes at wjec.co.uk. Always revise from the current specification and WJEC's own past papers, because question style is board-specific.

Sources & how we know this

WJEC GCE AS/A level Computer Science specification (from 2015) — WJEC (2015)